Fascicle localisation within peripheral nerves through evoked activity recordings: A comparison between electrical impedance tomography and multi-electrode arrays

Closed-Loop Control of Functional Electrical Stimulation Using a Selectively Recording and Bidirectional Nerve Cuff Interface

Discriminating recorded afferent neural information can provide sensory feedback for closed-loop control of functional electrical stimulation, which restores movement to paralyzed limbs. Previous work achieved state-of-the-art off-line classification of electrical activity in different neural pathways recorded by a multi-contact nerve cuff electrode, by applying deep learning to spatiotemporal neural patterns. The objective of this study was to demonstrate the feasibility of this approach in the context of closed-loop stimulation. Acute in vivo experiments were conducted on 11 Long Evans rats to demonstrate closed-loop stimulation. A 64-channel ( 8×8 ) nerve cuff electrode was implanted on each rat's sciatic nerve for recording and stimulation. A convolutional neural network (CNN) was trained with spatiotemporal signal recordings associated with 3 different states of the hindpaw (dorsiflexion, plantarflexion, and pricking of the heel). After training, firing rates were reconstructed from the classifier outputs for each of the three target classes. A rule-based closed-loop controller was implemented to produce ankle movement trajectories using neural stimulation, based on the classified nerve recordings. Closed-loop stimulation was successfully demonstrated in 6 subjects. The number of successful movement sequence trials per subject ranged from 1-17 and number of correct state transitions per trial ranged from 3-53. This work demonstrates that a CNN applied to multi-contact nerve cuff recordings can be used for closed-loop control of functional electrical stimulation.

A scoping review of current and emerging techniques for evaluation of peripheral nerve health, degeneration, and regeneration: part 1, neurophysiology

Peripheral neuroregeneration research and therapeutic options are expanding exponentially. With this expansion comes an increasing need to reliably evaluate and quantify nerve health. Valid and responsive measures that can serve as biomarkers of the nerve status are essential for both clinical and research purposes for diagnosis, longitudinal follow-up, and monitoring the impact of any intervention. Furthermore, such biomarkers can elucidate regeneration mechanisms and open new avenues for research. Without these measures, clinical decision-making falls short, and research becomes more costly, time-consuming, and sometimes infeasible. As a companion to Part 2, which is focused on non-invasive imaging, Part 1 of this two-part scoping review systematically identifies and critically examines many current and emerging neurophysiological techniques that have the potential to evaluate peripheral nerve health, particularly from the perspective of regenerative therapies and research.

Deep-learning segmentation of fascicles from microCT of the human vagus nerve

Introduction

MicroCT of the three-dimensional fascicular organization of the human vagus nerve provides essential data to inform basic anatomy as well as the development and optimization of neuromodulation therapies. To process the images into usable formats for subsequent analysis and computational modeling, the fascicles must be segmented. Prior segmentations were completed manually due to the complex nature of the images, including variable contrast between tissue types and staining artifacts.

Methods

Here, we developed a U-Net convolutional neural network (CNN) to automate segmentation of fascicles in microCT of human vagus nerve.

Results

The U-Net segmentation of ~500 images spanning one cervical vagus nerve was completed in 24 s, versus ~40 h for manual segmentation, i.e., nearly four orders of magnitude faster. The automated segmentations had a Dice coefficient of 0.87, a measure of pixel-wise accuracy, thus suggesting a rapid and accurate segmentation. While Dice coefficients are a commonly used metric to assess segmentation performance, we also adapted a metric to assess fascicle-wise detection accuracy, which showed that our network accurately detects the majority of fascicles, but may under-detect smaller fascicles.

Discussion

This network and the associated performance metrics set a benchmark, using a standard U-Net CNN, for the application of deep-learning algorithms to segment fascicles from microCT images. The process may be further optimized by refining tissue staining methods, modifying network architecture, and expanding the ground-truth training data. The resulting three-dimensional segmentations of the human vagus nerve will provide unprecedented accuracy to define nerve morphology in computational models for the analysis and design of neuromodulation therapies.

Fibers in smaller fascicles have lower activation thresholds with cuff electrodes due to thinner perineurium and smaller cross-sectional area

Objective. In nerve stimulation therapies, fibers in larger fascicles generally have higher activation thresholds, but the mechanisms are not well understood. We implemented and analyzed computational models to uncover the effects of morphological parameters on activation thresholds.Approach. We implemented finite element models of human vagus nerve stimulation to quantify the effects of morphological parameters on thresholds in realistic nerves. We also implemented simplified models to isolate effects of perineurium thickness, endoneurium diameter, fiber diameter, and fascicle location on current density, potential distributions (Ve), and activation thresholds across cuff geometries and stimulation waveforms. UsingVefrom each finite element model, we simulated activation thresholds in biophysical cable models of mammalian axons.Main results. Perineurium thickness increases with fascicle diameter, and both thicker perineurium and larger endoneurial diameter contributed to higher activation thresholds via lower peak and broader longitudinal potentials. Thicker perineurium caused less current to enter the fascicle transversely, decreasing peakVe. Thicker perineurium also inhibited current from leaving the fascicle, causing more constant longitudinal current density, broadeningVe. With increasing endoneurial diameter, intrafascicular volume increased faster than surface area, thereby decreasing intrafascicular current density and peakVe. Additionally, larger fascicles have greater cross-sectional area, thereby facilitating longitudinal intrafascicular current flow and broadeningVe. A large neighboring fascicle could increase activation thresholds, and for a given fascicle, fiber diameter had the greatest effect on thresholds, followed by fascicle diameter, and lastly, fascicle location within the epineurium. The circumneural cuff elicited robust activation across the nerve, whereas a bipolar transverse cuff with small contacts delivering a pseudo-monophasic waveform enabled more selective activation across fiber diameters and locations.Significance. Our computational studies provide mechanistic understanding of neural responses across relevant morphological parameters of peripheral nerves, thereby informing rational design of effective therapies.

A combined cuff electrode array for organ-specific selective stimulation of vagus nerve enabled by Electrical Impedance Tomography

Previously developed spatially-selective Vagus Nerve Stimulation (sVNS) allows the targeting of specific nerve fascicles through current steering in a multi-electrode nerve cuff but relies on a trial-and-error strategy to identify the relative orientation between electrodes and fascicles. Fast Neural Electrical Impedance Tomography (FN-EIT) has been recently used for imaging neural traffic in the vagus nerves of pigs in a cross-correlation study with sVNS and MicroCT fascicle tracking. FN-EIT has the potential for allowing targeted sVNS; however, up to now, stimulation and imaging have been performed with separate electrode arrays. In this study, different options were evaluated in-silico to integrate EIT and stimulation into a single electrode array without affecting spatial selectivity. The original pig vagus EIT electrode array geometry was compared with a geometry integrating sVNS and EIT electrodes, and with direct use of sVNS electrodes for EIT imaging. Modelling results indicated that both new designs could achieve image quality similar to the original electrode geometry in all tested markers (e.g., co-localisation error <100 µm). The sVNS array was considered to be the simplest due to the lower number of electrodes. Experimental results from testing evoked EIT imaging of recurrent laryngeal activity using electrodes from the sVNS cuff returned a signal-to-noise ratio similar to our previous study (3.9 ± 2.4 vs. 4.1 ± 1.5, N = 4 nerves from 3 pigs) and a lower co-localisation error (≈14% nerve diameter vs. ≈25%, N = 2 nerves from 2 pigs). Performing FN-EIT and sVNS on the same nerve cuff will facilitate translation to humans, simplify surgery and enable targeted neuromodulation strategies.

Organotopic organization of the porcine mid-cervical vagus nerve

Introduction

Despite detailed characterization of fascicular organization of somatic nerves, the functional anatomy of fascicles evident in human and large mammal cervical vagus nerve is unknown. The vagus nerve is a prime target for intervention in the field of electroceuticals due to its extensive distribution to the heart, larynx, lungs, and abdominal viscera. However, current practice of the approved vagus nerve stimulation (VNS) technique is to stimulate the entire nerve. This produces indiscriminate stimulation of non-targeted effectors and undesired side effects. Selective neuromodulation is now a possibility with a spatially-selective vagal nerve cuff. However, this requires the knowledge of the fascicular organization at the level of cuff placement to inform selectivity of only the desired target organ or function.

Methods and results

We imaged function over milliseconds with fast neural electrical impedance tomography and selective stimulation, and found consistent spatially separated regions within the nerve correlating with the three fascicular groups of interest, suggesting organotopy. This was independently verified with structural imaging by tracing anatomical connections from the end organ with microCT and the development of an anatomical map of the vagus nerve. This confirmed organotopic organization.

Discussion

Here we show, for the first time, localized fascicles in the porcine cervical vagus nerve which map to cardiac, pulmonary and recurrent laryngeal function (N = 4). These findings pave the way for improved outcomes in VNS as unwanted side effects could be reduced by targeted selective stimulation of identified organ-specific fiber-containing fascicles and the extension of this technique clinically beyond the currently approved disorders to treat heart failure, chronic inflammatory disorders, and more.

Mapping the functional anatomy and topography of the cardiac autonomic innervation for selective cardiac neuromodulation using MicroCT

Background: Vagus nerve stimulation (VNS) has gained great importance as a promising therapy for a myriad of diseases. Of particular interest is the therapy of cardiovascular diseases, such as heart failure or atrial fibrillation using selective cardiac VNS. However, there is still a lack of organ-specific anatomical knowledge about the fascicular anatomy and topography of the cardiac branch (CB), which diminishes the therapeutic possibilities for selective cardiac neuromodulation. Here, we established a topographical and anatomical map of the superior cardiac VN in two animal species to dissect cervical and cardiac VN morphology.

Methods: Autonomic nerves including superior CBs were harvested from domestic pigs and New Zeeland rabbits followed by imaging with microcomputed tomography (µCT) and 3D rendering. The data were analyzed in terms of relevant topographical and anatomical parameters.

Results: Our data showed that cardiac vagal fascicles remained separated from other VN fascicles up to 22.19 mm (IQR 14.02–41.30 mm) in pigs and 7.68 mm (IQR 4.06–12.77 mm) in rabbits from the CB point and then started merging with other fascicles. Exchanges of nerve fascicles between sympathetic trunk (ST) and VN were observed in 3 out of 11 nerves, which might cause additional unwanted effects in unselective VNS. Our 3D rendered digital model of the cardiac fascicles was generated showing that CB first remained on the medial side where it branched off the VN, as also shown in the µCT data of 11 pig nerves, and then migrated towards the ventromedial site the further it was traced cranially.

Conclusion: Our data provided an anatomical map of the cardiac vagal branches including cervical VN and ST for future approaches of selective cardiac neurostimulation, indicating the best position of selective cardiac VNS just above the CB point.

Closed-Loop Vagus Nerve Stimulation for the Treatment of Cardiovascular Diseases: State of the Art and Future Directions

The autonomic nervous system exerts a fine beat-to-beat regulation of cardiovascular functions and is consequently involved in the onset and progression of many cardiovascular diseases (CVDs). Selective neuromodulation of the brain-heart axis with advanced neurotechnologies is an emerging approach to corroborate CVDs treatment when classical pharmacological agents show limited effectiveness. The vagus nerve is a major component of the cardiac neuroaxis, and vagus nerve stimulation (VNS) is a promising application to restore autonomic function under various pathological conditions. VNS has led to encouraging results in animal models of CVDs, but its translation to clinical practice has not been equally successful, calling for more investigation to optimize this technique. Herein we reviewed the state of the art of VNS for CVDs and discuss avenues for therapeutic optimization. Firstly, we provided a succinct description of cardiac vagal innervation anatomy and physiology and principles of VNS. Then, we examined the main clinical applications of VNS in CVDs and the related open challenges. Finally, we presented preclinical studies that aim at overcoming VNS limitations through optimization of anatomical targets, development of novel neural interface technologies, and design of efficient VNS closed-loop protocols.

Simplifying the hardware requirements for fast neural EIT of peripheral nerves

OBJECTIVE

The main objective of this study was to assess the feasibility of lowering the hardware requirements for fast neural EIT in order to support the distribution of this technique. Specifically, the feasibility of replacing the commercial modules present in the existing high-end setup with compact and cheap customized circuitry was assessed.

APPROACH

Nerve EIT imaging was performed on rat sciatic nerves with both our standard ScouseTom setup and a customized version in which commercial benchtop current sources were replaced by custom circuitry. Electrophysiological data and images collected in the same experimental conditions with the two setups were compared. Data from the customized setup was subject to a down-sampling analysis to simulate the use of a recording module with lower specifications.

MAIN RESULTS

Compound action potentials (573±287µV and 487±279µV, p=0.28) and impedance changes (36±14µV and 31±16µV, p=0.49) did not differ significantly when measured using commercial high-end current sources or our custom circuitry, respectively. Images reconstructed from both setups showed neglibile (<1voxel, i.e. 40µm) difference in peak location and a high degree of correlation (R2=0.97). When down-sampling from 24 to 16 bits ADC resolution and from 100KHz to 50KHz sampling frequency, signal-to-noise ratio showed acceptable decrease (<-20%), and no meaningful image quality loss was detected (peak location difference <1voxel, pixel-by-pixel correlation R2=0.99).

SIGNIFICANCE

The technology developed for this study greatly reduces the cost and size of a fast neural EIT setup without impacting quality and thus promotes the adoption of this technique by the neuroscience research community.

Computational modelling of nerve stimulation and recording with peripheral visceral neural interfaces

Objective.Neuromodulation of visceral nerves is being intensively studied for treating a wide range of conditions, but effective translation requires increasing the efficacy and predictability of neural interface performance. Here we use computational models of rat visceral nerve to predict how neuroanatomical variability could affect both electrical stimulation and recording with an experimental planar neural interface.Approach.We developed a hybrid computational pipeline,VisceralNerveEnsembleRecording andStimulation (ViNERS), to couple finite-element modelling of extracellular electrical fields with biophysical simulations of individual axons. Anatomical properties of fascicles and axons in rat pelvic and vagus nerves were measured or obtained from public datasets. To validate ViNERS, we simulated pelvic nerve stimulation and recording with an experimental four-electrode planar array.Main results.Axon diameters measured from pelvic nerve were used to model a population of myelinated and unmyelinated axons and simulate recordings of electrically evoked single-unit field potentials (SUFPs). Across visceral nerve fascicles of increasing size, our simulations predicted an increase in stimulation threshold and a decrease in SUFP amplitude. Simulated threshold changes were dominated by changes in perineurium thickness, which correlates with fascicle diameter. We also demonstrated that ViNERS could simulate recordings of electrically-evoked compound action potentials (ECAPs) that were qualitatively similar to pelvic nerve recording made with the array used for simulation.Significance.We introduce ViNERS as a new open-source computational tool for modelling large-scale stimulation and recording from visceral nerves. ViNERS predicts how neuroanatomical variation in rat pelvic nerve affects stimulation and recording with an experimental planar electrode array. We show ViNERS can simulate ECAPS that capture features of our recordings, but our results suggest the underlying NEURON models need to be further refined and specifically adapted to accurately simulate visceral nerve axons.